How to find lookalike customers from your best wins
When you sell into a new market, you have to describe who to go after. Most teams do it with filters: industry, headcount, country. The trouble is that a filter describes thousands of companies, and most of them will never buy.
There is a better place to start: the companies that already said yes. Find more companies like them, and you have lookalike customers (most tools call them lookalike companies). A filter starts from a guess. Your closed-won list starts from proof, because someone paid.
What a filter misses
Take a made-up company that sells route-planning software to food distributors. Its three biggest wins this year are a frozen-food distributor in Chicago, a dairy cold-chain company in London and a bakery-ingredients wholesaler in Sydney.
A filter would say "food distributors, 200 to 500 staff". The three wins say much more. All three run their own delivery fleet. All three hired a new Head of Operations in the six months before they bought. All three were hiring dispatchers at the time. Two had just raised money to deliver in new cities. Nobody would put that in a filter.
Study why your best customers bought when they did
A company does not become a good customer at random. Something changed that made the problem urgent, and made that quarter the moment to buy.
So take your best wins one at a time and ask what happened just before the deal. Who joined? Which team grew? Did they raise money, and what for? Did they enter a new market? Then put the answers side by side. The secret lies in the similarities. One win with a new Head of Operations is a coincidence. Three is a pattern you can search for.
The similarities worth looking for
| Similarity | What it can tell you | In the example |
|---|---|---|
| A senior hire | New leaders often review the tools and suppliers they inherit. | A new Head of Operations |
| A junior team growing fast in one function | The company is putting money behind that function now. | Several dispatcher roles open |
| A round raised for a stated purpose | The money already has a job, and the purpose says which one. | Funding to deliver in new cities |
| A specific region | Companies in the same market face the same costs and competitors. | Big cities in English-speaking markets |
| A team of a certain size | The problem only starts to hurt past a certain size. | A dispatch team of 10 to 30 |
| The size of one team against the whole company | It shows how the company really runs, whatever its website says. | Drivers are over half the staff |
| The tools they already use | Your product fits in without a fight. | A fleet-tracking system the software plugs into |
Websites that say the same things
Tools like Ocean.io take your best wins, read their websites, and return companies whose websites describe similar things, ranked. They are a fast way to cast a wide first net. In one of our tests, five starting companies, searched in one country, gave back 2,407 lookalikes. About one in three of the first 75 was new to us and a good fit.
That is why it is worth using. But a tool like this judges a company only by what its website says. Two companies can say the same things and be nothing alike. A food distributor and a food manufacturer both write about cold storage, freshness and delivery. Only the distributor plans delivery routes all day.
Or take two SaaS products that solve the same problem. One has five employees and is based in the Philippines. The other has 75 employees in New York and has just raised a Series B. Its sales team is mature: a VP of Sales, Sales Directors, SDR Managers, SDRs and AEs. Their websites read almost the same. They are not the same buyer.
And a perfect company whose website says very little never shows up at all.
Combine the signals to find lookalike customers
Each signal on its own is weak. Plenty of companies hire a new Head of Operations and never buy route-planning software. A website match can be a lucky choice of words. But a company that shows three or four of the signals your wins showed is probably a real fit.
So put every company in one list and count how many signals each one shows. Here is the made-up example again.
| Company | New Head of Operations | Hiring dispatchers | Drivers over half the staff | Website match | Signals |
|---|---|---|---|---|---|
| A | yes | yes | yes | yes | 4 |
| B | yes | yes | yes | 3 | |
| C | yes | 1 | |||
| D | yes | yes | yes | 3 |
Company A goes first. Company C looks right on its website, but nothing else backs it up, so the match may be chance. Company D has no website match, but it shows three other signals. It may be a hidden gem whose website says very little. A website tool alone would have ranked C above D.
Try it with your own wins
Send us three companies you closed recently. Pick your biggest recent wins, or the ones you had to win. We will send back ten ideal-match companies that have the problem you solve. Send your three companies